Locking and other disadvantages of solid elements

[Anonymous]

I would like to compare a plate model with a solid model. In the description it is said that you should have three elements in thickness.

Is that recommended for stress-states that are analysed through the thickness and you could also leave one element in thickness or does it have to do with locking?

Do the elements in ASIM have any safety against locking?

I mean you can increase the integration order, but there are several other methods one can find in literature, like reduced integration etc.

[Anonymous]

Look into the use of midside notes on the solid elements.

[S.LI]

In ASIM MES, mean-dilatation, one kind of selective reduced integration method, is provided for brick and tetrahedron elements.

Also, as mentioned, higher order elements are helpful, such as mid-side nodes.

[Anonymous]

Okay, so the reduced integration is only available in MES?

And what about these three elements in thickness? Are they only recommended to get a better result through the thickness or is there another reason related to locking or whatever?

[S.LI · Accepted solution]

In both MES and LSS, "compatible mode" is provided to avoid shear locking in solid elements, such as tet and brick.

The selective reduced integration in MES is usually for relaxing volume locking.

All these are about solid elements.

For shell/plate elements, people are talking about locking, too. But locking in shell is about simulating thin shell with thick-shell formulations. Reduced integration method was an effective way before new techs get introduced into FEA.

[Anonymous]

Hi,

I came across this statement while looking up information concerning formulations of the solid elements.  Would you please elaborate on the statement "compatible modes are provided to avoid shear locking"?

Also, is the selective reduced integration used only for eight node bricks, or is it also being used in 20 node bricks?  How about degenerated elements obtained by collapsing bricks (wedges, or tetrahedra): is the selective reduced integration still in effect?

Thanks,

Petr